AI Applications Engineer / Fractional CTO / 15 Years Shipping
BRINGING YOU
RETURN ON AI
INVESTMENT.
2009 / Toronto / Remote
AI Agents & MCP / Full-Stack Ownership / Team Enablement
01 / The Shift
The engineers who win next are the ones configuring the machine that builds the app.
Software engineering is undergoing a paradigm shift. The craft is moving from writing every line of code to defining systems so precisely that AI agents build them for you — faster, with fewer bugs, at scale.
I operate at this intersection. Spend more time defining and architecting so AI builds faster and smoother. Context infrastructure becomes the new codebase. PRDs become executable specifications. MCP (Model Context Protocol) is the open standard that lets AI agents connect to any external tool or data source — think of it as USB-C for AI. I build both sides of that protocol.
The teams I enable don't just use AI tools. They run automated pipelines where agents pull, branch, fix, push, and create PRs for human review. Engineers self-report 2-3x capacity and carry these practices to every future role.
“Not writing code for applications — configuring agentic systems that build them.”
Building AI Products
The output is AI — agents, MCP servers, LLM integrations, context-engineered tools.
Building With AI
The process uses AI — Claude Code setup, AI-first workflows, context infrastructure, PRD-driven development, agent sandboxes.
- Weekly Claude Code sync calls across the team
- Context infrastructure maintained for autonomous agents
- PRD-driven development with approval gates pre-build
- Automated pipelines: agents pull, branch, fix, push, PR
- Engineers report 2-3x capacity gains
02 / Credentials
Years Building & Shipping Software
From mobile to full-stack to AI-native — now building agents, MCP servers, and enabling teams to ship 2-3x faster with AI-first workflows.
5x
$1.3M
50+
3
University of Toronto / Computer Engineering
Technical Stack
- Claude Code
- MCP (Full Spec)
- LLM Integration (Claude API)
- Agents (Python & TypeScript)
- Context Engineering
- Agentic Frameworks
- Agent Sandboxes
- Automated Bug-Fix Pipelines
- GitHub Actions + AI
- PRD-Driven AI Development
- n8n Orchestration
- NestJS
- Next.js 14
- GraphQL (Apollo)
- PostgreSQL
- DynamoDB
- Supabase
- Prisma
- Elasticsearch
- React 18
- TypeScript
- Tauri 2
- Rust
- Tailwind CSS
- Vite
- shadcn/ui
03 / Core Capabilities
Own Your Product
15 years of full-stack ownership. Give me a product — mobile, web, backend, infra — I'll own the whole thing. 3 Fractional CTO engagements. $1.3M raised.
- Fraq — solo-built SaaS: mobile + web + GraphQL API, deploying v2
- Coretext — CTO, hired 3 engineers, shipped near-complete platform in 6 months
- Homi — coded iOS + Android, 40+ device integrations, raised $1.3M seed
Build AI Agents
Custom agents deployed in production — financial services, investment banking, workforce management. Not prototypes. Agents that run businesses.
- Client onboarding agent — scaled throughput 5x (Hong Kong, financial services)
- Market research agent — live data + proprietary strategy testing (investment banking)
- Workforce management tool — 5 locations, auto-enrichment (Colombia)
Ship MCP Servers
Full-spec MCP development across transports, primitives, and auth. Both sides of the protocol — servers and clients.
- SVG Development Server — AI coding agents working with design assets natively
- Flume AI Internal Server — ConstructionWire + Apollo + HubSpot, full OAuth
- Coretext Sidekick — built a desktop MCP client from scratch (Tauri 2 + Rust)
Enable AI-First Teams
Train your engineers on Claude Code. Set up context infrastructure, agentic pipelines, and PRD-driven workflows. Teams I've enabled report 2-3x velocity gains they carry to future roles.
- Coretext — 3 engineers, 2-3x capacity, practices carried forward
- Client team — 1-month engagement, codebase onboarded to AI-first, 2x velocity
- Automated bug-fix pipelines — agents pull, branch, fix, push, PR for human review
04 / Proof of Impact
AI Agents: Production Agent Development
Specialized agents deployed in cloud sandbox environments for financial services, investment banking, and workforce management. Each agent solves a specific business problem — not demos, not prototypes, running operations.
Active — ongoing client engagements
MCP Servers: Full-Spec Protocol Development
3 MCP servers in production spanning local and remote transports. Full feature set: tools, resources, prompts, sampling, OAuth 2.0. From AI coding agent tooling to enterprise workflow integration to QA automation with subagent spawning.
Active — SVG server open sourcing soon
Spanish Tutor Agent: Conversational Language Learning
AI-powered Spanish tutor that conducts real-time voice conversations, corrects grammar inline, adapts difficulty to learner level, and tracks progress across sessions. Voice-first interface with Whisper STT and ElevenLabs TTS for natural conversation practice.
Active — in development
Knowledge Graph Engine: Graph-Based Retrieval (RAG Alternative)
A knowledge graph system that replaces traditional vector-based RAG with structured graph traversal. Documents are decomposed into entities and relationships, stored in Neo4j, and queried via LLM-generated Cypher — producing more accurate, explainable, and context-rich retrievals than embedding similarity.
Active — in development
Microsoft MCP Server: Enterprise Integration on Azure
Production MCP server built on the Microsoft stack — .NET 8, Azure Functions, Microsoft Graph API. Gives AI agents structured access to Teams, SharePoint, Outlook, OneDrive, and Planner. Entra ID authentication with granular permission scoping. Deployed as serverless Azure Functions.
Active — in development
Video Ad Pipeline: AI-Generated Video Advertising
End-to-end pipeline for generating custom video advertisements from text briefs. Takes a creative brief, generates script and storyboard, produces video segments via AI, synthesizes voiceover, composites with motion graphics, and delivers review-ready assets. Orchestrated with n8n, review dashboard in Next.js.
Active — in development
Coretext: AI Tooling Platform
Fractional CTO for AI tooling startup. Built team and product from zero in 6 months. Sidekick: desktop MCP client (Tauri 2 + Rust). Console: web app for configuring custom MCP servers. Enabled team with AI-first practices — engineers reported 2-3x capacity. Releasing Sidekick as open source.
Open sourcing Sidekick
Fraq: Fractional Ownership Platform
Solo-built the entire tech stack for a fractional vacation property platform across Latin America. Cross-platform mobile app, web dashboards, GraphQL API, and complex business logic — token economics, multi-currency, role-based access for 4 user roles.
Active — deploying v2.0
05 / Lab
Coretext Sidekick
Open SourcingDesktop MCP client built with Tauri 2 + Rust. Multi-workspace interface, multi-MCP server connections, real-time streaming chat, file management, dynamic theming. Retained all IP — releasing as standalone open source.
SVG Development MCP Server
Open Sourcing SoonInstructions, resources, guides, and tools for AI coding agents working with SVG assets natively. Built in collaboration with a design team. Enabling agents to create, edit, and reason about vector graphics.
Claude Code Workflows
OngoingContext infrastructure, agent configurations, team enablement patterns. CLAUDE.md files, hooks, custom commands, skills, and automated pipelines for AI-first development teams.
Obsidian-driven context management system for managing multiple projects across multiple clients. Focuses on context retrieval and assembly — the same context engineering practices used for AI agents, applied to human project management.
Curated Claude Code plugins for software development. Skills, hooks, and configurations that accelerate common development workflows — from code review to deployment automation.
Image generation skills for Claude Code. AI-powered visual asset creation integrated directly into development workflows — generate, edit, and optimize images without leaving the terminal.
Claude Code marketplace for Kotlin Multiplatform development. Skills, prompts, and configurations purpose-built for KMP projects — Compose Multiplatform, platform-specific code, and shared module patterns.
Claude Code
Primary development environment
MCP
Agent-to-tool protocol layer
n8n
Agent orchestration and automation
Tauri 2
Cross-platform desktop apps
Next.js 15
Web applications
Supabase
Backend and auth
Neo4j
Graph database
Azure
Enterprise cloud
06 / Career Trajectory
Mobile engineering at Samsung and Bose. Full-stack ownership at startups. Three fractional CTO roles. Now building the AI application layer. The constant: give me a product, I'll own it.
AI Applications Engineer / Consultant
SELF-EMPLOYED
Building AI agents, MCP servers, and enabling teams with AI-first development practices. Clients across financial services, investment banking, and workforce management. Training engineers on Claude Code. Setting up agentic frameworks for production codebases.
- Client onboarding agent — WhatsApp integration, scaled throughput 5x (Hong Kong, financial services)
- 3 MCP servers in production — full spec across stdio, HTTP+SSE, OAuth, and subagent spawning
- Onboarded engineering team to Claude Code and AI-first practices — 2x velocity in 1 month
- Market research agent with live market data for investment banking strategy testing
Fractional CTO
CORETEXT
Fractional CTO for AI tooling startup. Built a desktop MCP client (Sidekick) and MCP server configuration platform (Console) with a team of 3 engineers in 6 months. Implemented AI-first team practices that produced 2-3x velocity gains. Releasing Sidekick as open source.
- Implemented AI-first practices — weekly Claude Code syncs, context infrastructure, PRD-driven development
- Engineers self-reported 2-3x capacity vs. previous workflows — carrying practices to future roles
- Set up automated bug-fix pipelines — agents pull, branch, fix, push, create PR for human review
- Personally coded Sidekick desktop app (Tauri 2 + Rust + React + TypeScript)
Fractional CTO
FRAQ
Solo technical founder building the entire tech stack for a fractional vacation property ownership platform across Latin America. Property Management System serving admins, property managers, operators, and owners.
- Solo-built entire SaaS: mobile (KMP), web (Next.js 14), backend (GraphQL + Supabase)
- Complex business domains: token wallets, 3-year rolling windows, 7 currencies, 4 user roles
- Hired and managed 5-8 contractors across design and development
- Deploying v2.0 PMS + mobile apps now
Senior Android Developer
THOUGHTWORKS
Monetization team for a confidential North American virtual fitness platform. Defined and developed core subscription architecture including payments and tiered subscriptions.
- Defined & developed core subscription architecture
- Built parent fragment architecture with custom injections
- Data stores, fetchers, mappers, repositories
- Managed developers, facilitated knowledge transfer
Fractional CTO
HOMI SMARTHOME
The First Smart Home Managed Service in Asia. Defined architecture for fully remote-connected smart home system (v2). Went to Singapore with CEO and raised $1.3M seed round.
- Raised $1.3M seed round (SeedPlus + others)
- Defined v2 architecture transforming local-only to fully remote smart home system
- Coded iOS and Android mobile applications with 40+ smart device integrations
- Hired IoT cloud development team and QA/servicing engineers
Senior Android Developer
VARIOUS
Contract engineering across major brands including Samsung, Bose, The RealReal, Univision, and RBC. Shipped consumer-facing Android apps at scale.
- Samsung Galaxy Edge — VR experiences and edge features
- Bose (BoseConnect) — Reactive interfaces, TDD, Material Design
- The RealReal — Sole Android developer, full ownership through release cycles
- RBC Mobile — NFC payments and SIM-based security
07 / Let's Work Together
Embedded AI Engineer
Join your team for 3-6 months. Build agents, MCP servers, and AI infrastructure alongside your engineers.
Full integration with your team's codebase, rituals, and roadmap
Fractional CTO
Own your technical strategy. Architecture, hiring, roadmap, and hands-on coding. 3 engagements and counting.
From zero to shipped product with a team I build and manage
Team Enablement
1-month sprint. Claude Code setup, context infrastructure, AI-first workflows. Teams report 2-3x velocity.
Your engineers carry these practices to every future project
Advisory
Ongoing technical advisory. Architecture reviews, AI strategy, vendor evaluation. Async or weekly sync.
Second opinion from someone who ships, not just advises
Currently accepting 1-2 new engagements